Influences of Process Parameters on Residual Stress and Tensile Performance of Riveted Joints

The mechanical integrity and performance of riveted joints critically influence the reliability and safety of automotive structures. To investigate the effect of residual stress on the tensile behavior of riveted joints, this study firstly established a finite element model of the riveting process using Abaqus software. The model was validated using the thick-walled cylinder theory, and the distribution of residual stress around the rivet hole was subsequently analyzed. Orthogonal experiments were subsequently conducted to examine the influence of key process parameters including rivet spacing, hole diameter, and upset head height on tensile performance, yielding simulation results for both radial residual stress and maximum tensile load. Based on the finite element analysis results, XGBoost and Kriging surrogate models were established using four key parameters as input variables, namely rivet spacing, hole diameter, upset head height, and arrangement pattern, with radial residual compressive stress and maximum tensile load as the output targets. Following a comparative evaluation of the two models' predictive accuracy, which revealed the superior performance of the Kriging model, it was subsequently integrated with the particle swarm optimization algorithm to conduct multi-objective process optimization. Results demonstrated that the optimal process parameters obtained through this approach increase the average residual compressive stress by 7.295% and the maximum tensile load by 11.237%, confirming significant performance improvement after process optimization.